Computer models flag rubidium-gold fluoride perovskites as lead-free candidates for solar and thermoelectric devices

A team of computational materials scientists has reported a detailed density functional theory study of two fluoride double perovskites, Rb₂AuCoF₆ and Rb₂AuRhF₆, finding that both compounds combine visible-range band gaps, strong light absorption and promising thermoelectric figures of merit. The work, published in Gold Bulletin on 27 November 2025, is entirely computational: no crystals were synthesized or measured, and the authors explicitly call for experimental synthesis and characterization to confirm their predictions.

The researchers, led by Wang Yurong with co-authors including Aparna Dixit, Jisha Annie Abraham, Preeti Kumari, Meznah M. Alanazi, Taoufik Saidani and Ramesh Sharma, chose this particular chemical family to address a persistent problem in perovskite research. Lead-containing halide perovskites deliver outstanding optoelectronic performance in solar cells, photodetectors and thermoelectric devices, but their toxicity and instability raise serious environmental concerns. Known lead-free alternatives such as Cs₂AgBiX₆ and Cs₂Au₂X₆ (X = Cl, Br, I) tend to have band gaps that are too wide and carrier mobility that is too low for efficient visible-light absorption and charge transport. The authors argue that replacing caesium with rubidium introduces greater lattice compactness and improved electronic coupling, while placing a transition metal such as cobalt or rhodium at the B-site allows band gap tuning and better charge transport through d–p orbital hybridization. According to their literature review, Rb₂AuMF₆ with M = Co or Rh had not previously been the subject of any theoretical investigation.

All calculations were performed with the WIEN2k code, which uses the full-potential linearized augmented plane wave method. The team employed the PBE generalized gradient approximation and the modified Becke-Johnson (mBJ) potential for the electronic structure, adding spin-orbit coupling in a further set of calculations to capture relativistic effects from the heavy gold and rubidium atoms. The primitive unit cell contains ten atoms per formula unit: two rubidium, one gold, one cobalt or rhodium, and six fluorine atoms. A k-mesh of 15 × 15 × 15 with 3000 k-points was used, and transport properties were evaluated with the BoltzTraP package based on semiclassical Boltzmann transport theory. Because scattering times are difficult to compute from first principles, the authors assumed a constant relaxation time of 10⁻¹⁴ seconds, and lattice thermal conductivity was estimated with Slack’s model rather than computed from phonons.

Structural screening came first. Both compounds adopt a cubic structure in the Fm-3̄m space group, with rubidium at (0.25, 0.25, 0.25), gold at (0, 0, 0.5) and cobalt or rhodium at (0, 0, 0). Energy-volume curves were fitted with the Birch-Murnaghan equation of state. The calculated bulk moduli are 66.72 GPa for Rb₂AuCoF₆ and 65.38 GPa for Rb₂AuRhF₆, values the authors describe as reflecting mechanical robustness. The lattice constant increases when cobalt is replaced by the larger rhodium atom, and the bulk modulus decreases accordingly. Negative formation energies and positive cohesive energies indicate that both compounds are thermodynamically more stable than their constituent elements, and both the Goldschmidt tolerance factor and the octahedral factor fall within the ranges conventionally associated with cubic stability, 0.813 to 1.107 and 0.415 to 0.895 respectively.

The electronic structure results place both compounds squarely in the visible-light regime. With the mBJ approximation, the band gaps fall between 1.454 and 1.828 electronvolts, a window the authors associate with potential suitability for solar energy harvesting. Including spin-orbit coupling moderately reduces the gaps to 1.269 eV for Rb₂AuCoF₆ and 1.592 eV for Rb₂AuRhF₆, while slightly modifying the band dispersion near the band edges. In the results and discussion section the band structures are described as indirect, with only slight variations in fine features and energy values between the two compounds. The partial density of states shows that gold d, fluorine p and cobalt d-t₂g orbitals dominate the valence band of Rb₂AuCoF₆, with analogous behaviour in the rhodium compound, while rubidium contributes negligibly near the Fermi energy. Charge density analysis indicates predominantly ionic bonding involving rubidium and covalent bonding between fluorine and the cobalt or rhodium site.

The optical analysis, also carried out with the mBJ potential, supports the photovoltaic motivation of the study. Reflectivity at zero photon energy is below 0.1 for both compounds, with the abstract citing a low reflectivity range of 0.05 to 0.08. Static refractive indices are 1.83 for Rb₂AuCoF₆ and 1.65 for Rb₂AuRhF₆, consistent with the inverse relationship between refractive index and band gap. The absorption coefficient peaks at 7.95 eV for the cobalt compound and 8.10 eV for the rhodium compound, and both materials show substantial absorption across 7.50 to 8.50 eV. The authors link this range to the most intense region of solar radiation and conclude that the compounds may be appropriate for sun absorbers, photovoltaics and other optoelectronic devices. Maximum reflectance values of 0.60 and 0.42 occur at 2.52 and 3.0 eV respectively, and the imaginary part of the dielectric function shows absorption edges that closely track the calculated band gaps.

The thermoelectric analysis is where the quantitative predictions become most detailed. The Seebeck coefficient is positive at room temperature, peaks near 200 K at roughly 181 microvolts per kelvin, and decreases progressively up to 1200 K for both compounds, behaviour the authors attribute to p-type conduction. Power factors rise over the 200 to 1200 K range, peaking at 450 K before declining, with maximum values of 5.87 × 10¹¹ W/mK²s for Rb₂AuCoF₆ and 6.40 × 10¹¹ W/mK²s for Rb₂AuRhF₆ at 600 K. Electrical conductivity, reported as the conductivity divided by the relaxation time, increases nearly linearly with temperature and reaches 2.63 × 10¹⁹ and 2.92 × 10¹⁹ Ω⁻¹m⁻¹s⁻¹ for the cobalt and rhodium compounds respectively at 1200 K. Electronic thermal conductivity mirrors this trend.

Lattice thermal conductivity, estimated with the Slack model, comes out at 0.60 W/mK for Rb₂AuCoF₆ and 0.89 W/mK for Rb₂AuRhF₆ at room temperature. The authors compare these values with well-known thermoelectric materials, including SnSe at 0.62 W/mK, nano-twinned PbTe at 0.4 W/mK and Bi₂Te₃ at 0.35 W/mK at 473 K, while cautioning that differences between computational approaches mean experimental studies are needed for confirmation. The dimensionless figure of merit ZT reaches 0.69 for Rb₂AuCoF₆ and 0.63 for Rb₂AuRhF₆ at 1200 K in the temperature-dependent analysis, while the abstract quotes ZT values of 0.177 and 0.138, and an electronic ZT of roughly 0.9 is reported in the chemical-potential analysis at 300 and 700 K within an energy window of −2.0 to 2.0 eV. In the p-type response at a chemical potential of 1 eV, values exceed 0.8 for Rb₂AuCoF₆, whereas the n-type response is absent at 300 K. The authors suggest Rb₂AuRhF₆ merits further study as a p-type thermoelectric material near room temperature, while Rb₂AuCoF₆ may be more relevant at elevated temperatures, and identify Rb₂AuCoF₆ as the comparatively more promising thermoelectric candidate of the pair.

The authors frame the results as a groundwork-laying contribution rather than a demonstration of working devices. They propose that experimental synthesis could proceed via solid-state or solvothermal routes under fluorine-rich conditions, followed by X-ray diffraction and Raman or FTIR spectroscopy for phase identification, UV-Vis absorption and photoluminescence spectroscopy to confirm optical band gaps, Hall effect and Seebeck measurements to evaluate carrier transport, and X-ray photoelectron spectroscopy to verify the oxidation states of gold and the transition metal cations. Until such measurements exist, every number in the study, from the band gaps to the figure of merit, remains a model prediction dependent on the chosen exchange-correlation treatment, the assumed constant relaxation time and the Slack-model estimate of lattice thermal conductivity. Within those limits, the study adds two lead-free fluoride double perovskites to the growing catalogue of computationally screened candidates for visible-light optoelectronics and thermoelectric energy conversion.

Subject of Research: Computational prediction of structural, optoelectronic and thermoelectric properties of the lead-free fluoride double perovskites Rb₂AuCoF₆ and Rb₂AuRhF₆ using density functional theory

Article Title: Exploration of thermoelectric and optoelectronic features Rb₂AuMF₆ (M = Co, Rh) double perovskites for advanced technological applications: a computational evaluation

Article References: Yurong, W., Dixit, A., Abraham, J. A., Kumari, P., Alanazi, M. M., Saidani, T., & Sharma, R. (2025). Exploration of thermoelectric and optoelectronic features Rb₂AuMF₆ (M = Co, Rh) double perovskites for advanced technological applications: a computational evaluation. Gold Bulletin, 58(2), Article 16. https://doi.org/10.1007/s13404-025-00369-7

Image Credits: AI Generated

DOI: 10.1007/s13404-025-00369-7

Keywords: double perovskites, Rb2AuCoF6, Rb2AuRhF6, density functional theory, thermoelectrics, band gap, optoelectronics, lead-free perovskites, BoltzTraP, solar energy harvesting, lattice thermal conductivity, computational materials science